CN107248296B - Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics - Google Patents
Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics Download PDFInfo
- Publication number
- CN107248296B CN107248296B CN201710571448.2A CN201710571448A CN107248296B CN 107248296 B CN107248296 B CN 107248296B CN 201710571448 A CN201710571448 A CN 201710571448A CN 107248296 B CN107248296 B CN 107248296B
- Authority
- CN
- China
- Prior art keywords
- video
- image
- traffic flow
- time
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30236—Traffic on road, railway or crossing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
- Traffic Control Systems (AREA)
Abstract
本发明公开了一种基于无人机和时序特征的视频交通流量统计方法，属于图像处理技术领域。本发明首先通过无人机在公路交叉路口上方拍摄得到车流量视频，对得到的视频进行图像匹配以消除抖动带来的影响，然后利用取中值法提取出背景图像。对于视频的每一帧图像，再次进行与背景图像进行匹配。然后将匹配的图像指定区域进行特征提取，得到特征时序直方图。对特征时序直方图进行分析，得到车道的车流量数。该方法简单高效地提取区域特征，进行多次的排除无人机抖动带来的干扰误差，对车流量的统计更加精确。此方法简单高效，适应性高，可扩展性强，具有很广阔的应用前景。
The invention discloses a video traffic flow statistics method based on unmanned aerial vehicles and time sequence features, and belongs to the technical field of image processing. The present invention firstly obtains the traffic flow video by shooting the drone above the highway intersection, performs image matching on the obtained video to eliminate the influence of jitter, and then uses the median method to extract the background image. For each frame of the video, match with the background image again. Then feature extraction is performed on the specified area of the matched image, and the feature time series histogram is obtained. The characteristic time series histogram is analyzed to obtain the number of traffic flow in the lane. The method extracts regional features simply and efficiently, and eliminates the interference error caused by drone jitter for many times, and the statistics of traffic flow are more accurate. This method is simple and efficient, has high adaptability, strong expansibility, and has broad application prospects.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a traffic flow statistical method based on unmanned aerial vehicles and time sequence characteristic video processing.
Background
Intelligent transportation has become a future development trend, and how to automatically and efficiently manage a transportation system is also a current hotspot. The traffic flow detection is used as a part of intelligent traffic, and has a very important position in the aspects of traffic monitoring management, urban road construction and the like.
The major traffic flow detection techniques in recent years include: magnetic induction detection technology, wave frequency detection technology, video detection technology and the like. The video detection technology has the advantages of flexible installation, low cost, convenient management and maintenance and the like, and along with the development of image processing technology and computer vision, the traffic flow detection technology based on video images has attracted more and more attention and attention of people.
The current commonly used video-based vehicle detection algorithms mainly include: background subtraction, frame-to-frame subtraction, edge detection, optical flow tracking, and the like. Compared with the traditional method, the method effectively completes the work of monitoring, controlling and managing the road traffic through image processing and machine vision technology. Still have some shortcomings, for example every lane all need install the camera, easily produce lou to examine, the false retrieval scheduling problem.
Disclosure of Invention
The purpose of the invention is as follows: in order to solve the existing defects and problems, the invention provides a flexible, convenient, efficient and accurate traffic flow detection method.
The technical scheme is as follows: in order to achieve the purpose, the invention adopts the technical scheme that:
a video traffic flow statistical method based on unmanned aerial vehicles and time sequence characteristics comprises the following steps:
step 1, vertically shooting and recording downwards above a road intersection through an unmanned aerial vehicle to obtain a traffic flow video of the intersection, and manually inputting the position of a lane needing to count the traffic flow in a video image;
step 2, carrying out video analysis processing on the traffic flow statistical video to obtain a background image without a vehicle;
step 3, extracting a frame of image of the video and preprocessing the frame of image to obtain an image matched with a background image so as to eliminate the influence caused by shaking during shooting;
step 4, extracting characteristic values of the preprocessed image at the positions of lanes needing to be counted, extracting the next frame of the video, and repeating the steps 3 to 4, wherein the characteristic values of all the video frames are put together to form a characteristic time sequence histogram;
and 5, analyzing the characteristic time sequence histogram to obtain the traffic flow number of the lane to be counted in the time period.
Further, the step 2 is realized according to the following method:
step 2-1, taking 30 video images at equal time intervals from the video counting start time to the counting end time, and taking the first image as a calibration image;
2-2, respectively matching the 29 subsequent images with the first calibration image for surf characteristic points to obtain 29 new images, and forming 30 preprocessed images together with the original first calibration image;
and 2-3, traversing the pixels of each position of the 30 images, taking the median of three channel color components of the 30 pixels of each position as the pixel value of the specified channel of the specified position, and obtaining the background image without the vehicle after traversing.
Further, the pretreatment method in step 3 is as follows: and carrying out surf feature point matching on the video frame image and the background image.
Further, the step 4 is realized according to the following method;
step 4-1, respectively traversing all pixels of the manually input lane position area for the background image and the video frame image;
step 4-2, respectively calculating the absolute values of the difference values of the three color components for the pixels at the corresponding positions of each pair of images;
step 4-3, adding the absolute values of the difference values of the three components to obtain a pixel difference value of the pixel position of the two images;
step 4-4, adding the pixel difference values of all pixels in the area to be used as the characteristic value of the lane area;
and 4-5, if the region characteristic value is smaller than a certain threshold value, determining that the characteristic value is caused by noise, and setting the region characteristic value to be 0.
Further, the corresponding threshold selected in the step 4-5 is 100.
Further, the step 5 is realized according to the following method;
step 5-1, selecting a response characteristic threshold according to the time sequence histogram;
step 5-2, recording the time when each pair of response characteristic values of the time sequence histogram exceed the threshold value from left to right and the time when the response characteristic values lower than the threshold value from left to right as possible response time of the vehicle;
step 5-3, if the vehicle driving response time is less than 5 frames, the response is considered as noise interference, and the response is eliminated;
and 5-4, counting the number of all effective responses, namely the number of vehicles passing through the specified lane in the specified time.
Further, the method for selecting the corresponding threshold in the step 5-1 includes: and finding out the maximum response threshold value appearing in all time sequences, and selecting the threshold value as 10% of the maximum response threshold value.
Has the advantages that: the video traffic flow statistical method based on the unmanned aerial vehicle and the time sequence characteristics, provided by the invention, adopts the unmanned aerial vehicle technology which is rapidly developed in recent years to collect traffic flow videos, the collection mode becomes flexible and efficient, the traffic flow conditions of all lanes can be recorded at a traffic intersection at one time, and the complex operation of installing cameras at a plurality of intersections is saved. For the problem of jitter caused by images shot by an unmanned aerial vehicle in the air, surf characteristic transformation in digital image processing is adopted for image matching, and jitter is effectively eliminated. For the statistics of the traffic flow, only manually input lane areas are counted, and the interference error of the area outside the lane is eliminated. And simple and efficient regional characteristic time sequence analysis is adopted, multiple interference error elimination steps are carried out, and the statistics of the traffic flow is more accurate. The method is simple and efficient, high in adaptability and expandability and has wide application prospects.
Drawings
Fig. 1 is an algorithm flow chart of a video traffic flow statistical method based on unmanned aerial vehicles and time sequence characteristics according to the invention.
Fig. 2 is an input video shot by the drone.
Fig. 3 is a schematic diagram of selecting a lane area position.
Fig. 4 is a background image obtained by the median method.
Fig. 5 is a feature timing histogram.
Detailed Description
The present invention will be further described with reference to the accompanying drawings.
The method comprises the steps of firstly, shooting a traffic flow video above a road intersection by an unmanned aerial vehicle, preprocessing the obtained video, and extracting a background image. To eliminate the jitter, for each frame of image of the video, a match is made with the background image. And then, performing feature extraction on the matched image designated area to obtain a feature time sequence histogram. And analyzing the characteristic time sequence histogram to obtain the traffic flow number of the lane. The flow chart is shown in fig. 1.
Examples
Step 1, collecting images and inputting statistical area information.
Step 1-1, vertically shooting and recording downwards above a road intersection through an unmanned aerial vehicle to obtain a traffic flow video of the intersection, as shown in fig. 2;
step 1-2, selecting the position of the lane needing to be counted in the video as an area condition input system needing to be counted, and generally selecting an area with relatively pure color from a white line of the lane to a crosswalk, as shown in fig. 3.
And 2, calculating and extracting a background image by a median method.
Step 2-1, taking 30 video images at equal time intervals from the video counting start time to the counting end time, and taking the first image as a calibration image;
2-2, as the unmanned aerial vehicle inevitably shakes during aerial video shooting, matching transformation is required to eliminate the shaking so as to avoid error interference caused by shaking, and surf characteristic point matching transformation is respectively carried out on the 29 subsequent images and the first calibrated image to obtain new 29 images, and the new 29 images and the original first calibrated image together form 30 images after preprocessing;
and 2-3, traversing the pixels of each position of the 30 images, taking the median of three channel color components of the 30 pixels of each position as the pixel value of the specified channel of the specified position, and obtaining a background image without vehicles after traversing, as shown in fig. 4.
And 3, processing each frame in the video, extracting each frame, and performing surf feature matching on the frame image and the calibration image to achieve the purpose of eliminating the influence caused by shaking.
And 4, comparing the positions of the input regions in the frame image and the background image, extracting the response of the vehicle possibility in the region, and forming a characteristic time sequence histogram by the responses of all the frames together, as shown in fig. 5.
Step 4-1, respectively traversing all pixels of the manually input lane position area for the background image and the video frame image;
step 4-2, respectively calculating the absolute values of the difference values of the three color components for the pixels at the corresponding positions of each pair of images;
step 4-3, adding the absolute values of the difference values of the three components to obtain a pixel difference value of the pixel position of the two images;
step 4-4, adding the pixel difference values of all pixels in the area to be used as the characteristic value of the lane area;
and 4-5, if the region characteristic value is less than 100, determining that the characteristic value is caused by noise, and setting the region characteristic value to be 0, otherwise, keeping the region characteristic value.
And 5, analyzing the characteristic time sequence histogram to obtain the traffic flow number of the lane to be counted in the time period.
Step 5-1, selecting a response characteristic threshold according to the time sequence histogram, and generally selecting 10% of the maximum value;
step 5-2, recording the time when each pair of response characteristic values of the time sequence histogram exceed the threshold value from left to right and the time when the response characteristic values lower than the threshold value from left to right as possible response time of the vehicle;
step 5-3, if the vehicle driving response time is less than 5 frames, the response is considered as noise interference, and the response is eliminated;
and 5-4, counting the number of all effective responses, namely the number of vehicles passing through the specified lane in the specified time.
In conclusion, the unmanned aerial vehicle and the digital image processing technology are combined, and a convenient, efficient and accurate traffic flow detection scheme is realized through methods of background modeling, image jitter elimination, feature extraction and the like.
The accuracy rate of the traffic flow statistics can reach more than 90%, compared with the traditional method, the method can greatly save the cost of statistics, including manpower and material resources, improve the statistics efficiency, and ensure high accuracy rate. Due to the flexibility and convenience of the unmanned aerial vehicle, the method is wide in application range, high in expandability and wide in application prospect.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and these are intended to be within the scope of the invention.
Claims (1)
1. A video traffic flow statistical method based on unmanned aerial vehicles and time sequence characteristics is characterized in that: comprises the following steps:
step 1, vertically shooting and recording downwards above a road intersection through an unmanned aerial vehicle to obtain a traffic flow video of the intersection, and manually inputting the position of a lane needing to count the traffic flow in a video image;
step 2, carrying out video analysis processing on the traffic flow statistical video to obtain a background image without a vehicle;
step 3, extracting a frame of image of the video and preprocessing the frame of image to obtain an image matched with a background image so as to eliminate the influence caused by shaking during shooting;
step 4, extracting characteristic values of the preprocessed image at the positions of lanes needing to be counted, extracting the next frame of the video, and repeating the steps 3 to 4, wherein the characteristic values of all the video frames are put together to form a characteristic time sequence histogram;
step 5, analyzing the characteristic time sequence histogram to obtain the traffic flow number of the lane to be counted in a time period;
the step 2 is realized according to the following method:
step 2-1, taking 30 video images at equal time intervals from the video counting start time to the counting end time, and taking the first image as a calibration image;
2-2, respectively matching the 29 subsequent images with the first calibration image for surf characteristic points to obtain 29 new images, and forming 30 preprocessed images together with the original first calibration image;
step 2-3, traversing the pixels of each position of 30 images, taking the median of three channel color components of 30 pixels of each position as the pixel value of the specified channel of the specified position, and obtaining a background image without vehicles after traversing;
the pretreatment method in the step 3 comprises the following steps: performing surf feature point matching on the video frame image and the background image;
the step 4 is realized according to the following method;
step 4-1, respectively traversing all pixels of the manually input lane position area for the background image and the video frame image;
step 4-2, respectively calculating the absolute values of the difference values of the three color components for the pixels at the corresponding positions of each pair of images;
step 4-3, adding the absolute values of the difference values of the three components to obtain a pixel difference value of the pixel position of the two images;
step 4-4, adding the pixel difference values of all pixels in the area to be used as the characteristic value of the lane area;
step 4-5, if the area characteristic value is smaller than a certain threshold value, the characteristic value is considered to be caused by noise, and the area characteristic value is set to be 0;
the corresponding threshold value selected in the step 4-5 is 100;
the step 5 is realized according to the following method;
step 5-1, selecting a response characteristic threshold according to the time sequence histogram;
step 5-2, recording the time when each pair of response characteristic values of the time sequence histogram exceed the threshold value from left to right and the time when the response characteristic values lower than the threshold value from left to right as possible response time of the vehicle;
step 5-3, if the vehicle driving response time is less than 5 frames, the response is considered as noise interference, and the response is eliminated;
step 5-4, counting the number of all effective responses, namely the number of vehicles passing through the specified lane in the specified time;
the method for selecting the corresponding threshold value in the step 5-1 comprises the following steps: and finding out the maximum response threshold value appearing in all time sequences, and selecting the threshold value as 10% of the maximum response threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710571448.2A CN107248296B (en) | 2017-07-13 | 2017-07-13 | Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710571448.2A CN107248296B (en) | 2017-07-13 | 2017-07-13 | Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107248296A CN107248296A (en) | 2017-10-13 |
| CN107248296B true CN107248296B (en) | 2020-04-24 |
Family
ID=60013826
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710571448.2A Active CN107248296B (en) | 2017-07-13 | 2017-07-13 | Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107248296B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108364466B (en) * | 2018-02-11 | 2021-01-26 | 金陵科技学院 | Traffic flow statistical method based on unmanned aerial vehicle traffic video |
| CN108320510B (en) * | 2018-04-03 | 2020-12-04 | 深圳市智绘科技有限公司 | Traffic information statistical method and system based on aerial video shot by unmanned aerial vehicle |
| CN110689720A (en) * | 2019-10-10 | 2020-01-14 | 成都携恩科技有限公司 | Real-time dynamic traffic flow detection method based on unmanned aerial vehicle |
| CN112329631A (en) * | 2020-11-05 | 2021-02-05 | 浙江点辰航空科技有限公司 | Method for carrying out traffic flow statistics on expressway by using unmanned aerial vehicle |
| CN112542050B (en) * | 2020-12-23 | 2024-03-22 | 重庆市市政设计研究院有限公司 | Complex interchange traffic analysis device and method based on unmanned oblique photography |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001069569A3 (en) * | 2000-03-15 | 2002-01-31 | Raytheon Co | Automatic incident detection |
| CN103348380A (en) * | 2011-02-10 | 2013-10-09 | 日本电气株式会社 | Difference area detection system and difference area detection method |
| CN106570499A (en) * | 2016-10-28 | 2017-04-19 | 南京航空航天大学 | Object tracking method based on probability graph model |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080100473A1 (en) * | 2006-10-25 | 2008-05-01 | Siemens Corporate Research, Inc. | Spatial-temporal Image Analysis in Vehicle Detection Systems |
| CN100454355C (en) * | 2006-11-30 | 2009-01-21 | 复旦大学 | A video method for real-time collection of road traffic flow information |
| CN102324183B (en) * | 2011-09-19 | 2013-10-30 | 华中科技大学 | Method for detecting and shooting vehicle based on composite virtual coil |
| CN103413444B (en) * | 2013-08-26 | 2015-08-19 | 深圳市川大智胜科技发展有限公司 | A kind of traffic flow based on unmanned plane HD video is investigated method |
| CN105528891A (en) * | 2016-01-13 | 2016-04-27 | 深圳市中盟科技有限公司 | Traffic flow density detection method and system based on unmanned aerial vehicle monitoring |
-
2017
- 2017-07-13 CN CN201710571448.2A patent/CN107248296B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2001069569A3 (en) * | 2000-03-15 | 2002-01-31 | Raytheon Co | Automatic incident detection |
| CN103348380A (en) * | 2011-02-10 | 2013-10-09 | 日本电气株式会社 | Difference area detection system and difference area detection method |
| CN106570499A (en) * | 2016-10-28 | 2017-04-19 | 南京航空航天大学 | Object tracking method based on probability graph model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107248296A (en) | 2017-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107248296B (en) | Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics | |
| CN107315095B (en) | Multi-vehicle automatic speed measurement method with illumination adaptability based on video processing | |
| CN110660222B (en) | Intelligent environment-friendly electronic snapshot system for black-smoke road vehicle | |
| CN107133969B (en) | A kind of mobile platform moving target detecting method based on background back projection | |
| CN112347814B (en) | Passenger flow estimation and display method, system and computer-readable storage medium | |
| CN109145708B (en) | Pedestrian flow statistical method based on RGB and D information fusion | |
| CN106781680B (en) | A kind of curb parking intelligent control method based on the detection of image empty parking space | |
| CN108039044B (en) | System and method for intelligent queuing of vehicles based on multi-scale convolutional neural network | |
| CN102426785B (en) | Traffic flow information perception method and system based on contour and local feature points | |
| CN106951821A (en) | A method for intelligent monitoring and recognition of smoky vehicles based on image processing technology | |
| CN104778846A (en) | Computer-vision-based traffic light control method | |
| CN107590486B (en) | Moving object identification method and system, and bicycle flow statistical method and equipment | |
| CN104835147A (en) | Method for detecting crowded people flow in real time based on three-dimensional depth map data | |
| CN109684986B (en) | Vehicle analysis method and system based on vehicle detection and tracking | |
| WO2014082480A1 (en) | Method and device for calculating number of pedestrians and crowd movement directions | |
| CN103530638A (en) | Method for matching pedestrians under multiple cameras | |
| CN105321342A (en) | Intersection vehicle queuing length detection method based on aerial video | |
| CN102194270B (en) | Statistical method for pedestrian flow based on heuristic information | |
| CN109993134A (en) | A vehicle detection method at road intersection based on HOG and SVM classifier | |
| CN117911965B (en) | Expressway traffic accident identification method and equipment based on aerial image | |
| CN102663359B (en) | Method and system for pedestrian retrieval based on internet of things | |
| Brahme et al. | An implementation of moving object detection, tracking and counting objects for traffic surveillance system | |
| CN109584217B (en) | Automatic judging method for monitoring camera lens pollution | |
| CN104282157A (en) | Main line video traffic detecting method for traffic signal control | |
| CN109409264B (en) | Video queuing overflow detection method based on local tracking |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |